[Webinar] Beijing's Nuclear Evolution and China-U.S. Relations

Q: What is the risk of either side having tech glitches and have no response time to understand that it was a glitch and not a real firing on either side? IT systems are constantly crashing these days in all sectors of the global economy. Does this worry you?

 

Nuclear weapons and the organizational and technical systems that support them can be characterized as complex—having so many interactive parts that unpredictable failures can occur—and tightly coupled—components in one part of the system quickly and significantly impact components in other parts of the system.

 

The recently released film, A House of Dynamite, vividly illustrates the various technical and operational challenges that even a “mature” nuclear-armed state like the United States might confront in the short time between warning of a possible incoming strike and the time when the strike is expected to land.

 

Although all nuclear forces might be vulnerable to accidents, the specific degree to and ways in which a given nuclear force is susceptible to “tech glitches” may vary. A state which emphasizes the role of nuclear weapons in its military strategy, which adopts a launch-on-warning posture, has little redundancy in its early warning systems, and fields a vulnerable strategic deterrent may be more susceptible.

 

In designing their nuclear forces, including command and control structures, nuclear weapon states may face tradeoffs between enhancing the survivability of delivery systems and enhancing the survivability of command and control structures or between increasing resilience to cyberattacks and increasing resilience to kinetic attacks. There have been several “close calls” throughout history in which human or technical error generated false warnings of nuclear attack. 

​

 

Q: What is your view on the likelihood that the People's Republic of China (PRC) will adopt AI tools to bolster the speed of its possible response to a perceived nuclear first strike?

 

Difficult to say, though I don’t think this is likely. In 2024, U.S. President Joe Biden and Chinese President Xi Jinping met and, according to a readout of the discussion, “The two leaders affirmed the need to maintain human control over the decision to use nuclear weapons.”

 

Chinese strategists have generally been overly sanguine about the ability to avoid nuclear escalation. Some of my recent research also suggests that some Chinese strategists overlook the ways that incorporating artificial intelligence into nuclear-related operations could introduce or increase risks. However, China has historically prioritized strong centralized control over its nuclear and other non-nuclear strategic forces, suggesting Chinese decision-makers would be reluctant to incorporate AI tools into those processes.

 

The one area of strategic nuclear operations where decision-makers might think they could experience gains from incorporating artificial intelligence is in the detection and analysis of indications of an incoming strike. China is building the capabilities to support a launch-on-warning posture. If it adopted such a posture, it might depend on space-based infrared systems to detect ballistic missile launches and ground-based radars to track them in order to gain advance warning of an incoming strike and potentially launch before the strike arrived. Chinese strategists might (falsely) believe that AI could help avoid false negatives in these systems.

​

 

Q: Would you please elaborate more on the conventional and nuclear entanglement? Also, does the People's Liberation Army (PLA) have a concept like the US’s Conventional Support of Nuclear Operations (CSNO) or NATO’s Support of Nuclear Operations with Conventional Air Tactics (SNOWCAT)?

 

Conventional-nuclear entanglement refers to a range of circumstances in which the operations of nuclear forces may overlap with those of conventional forces. It can occur on three dimensions: geographic, operational, and technical.

 

Geographic entanglement occurs when conventional and nuclear forces are garrisoned or operate in the same spaces. Operational entanglement occurs when conventional and nuclear forces are operated by or rely on the same military institutions or practices. Technological entanglement occurs when conventional and nuclear forces are technically identical or indistinguishable.

 

In the People’s Liberation Army Rocket Force (PLARF), we see degrees of all three dimensions. The PLARF operates both conventional and nuclear ballistic and cruise missiles. Some of those systems might have blurring operational practices. And the PLARF fields a dual-capable intermediate range ballistic missile, the DF-26, that is supposedly “hot-swappable,” meaning it can change armament in the field. 

This entanglement can generate risks by possibly increasing the vulnerability of the entangled state’s nuclear forces and by increasing the possibility that conventional strikes by the entangled state are misperceived as nuclear strikes (or vice versa) or that strikes by another state targeting only the entangled conventional forces are instead meant to degrade the nuclear forces (or vice versa). 

 

My research shows that conventional-nuclear entanglement in the PLA Rocket Force may have not been a deliberate strategic choice but, rather, have emerged from pursuing narrower bureaucratic goals such as operational flexibility and cost savings. This, coupled with the fact that Chinese strategists are perhaps overconfident in their ability to avoid escalation to the nuclear threshold, could actually increase the possibility of dangerous misperception in a crisis or conflict.

​

 

Q: What are the prospects of the PRC changing its nuclear policy from "no first use" to a more aggressive posture?

 

Part of the answer depends on how one defines “no first use.” We have not seen any evidence that China is considering adopting a policy of deliberate first use. There are regular reports of elements of the PLA advocating for a less restrictive nuclear posture. The very existence of this kind of advocacy suggests that China’s "no first use" policy remains in place.

 

However, if China were to adopt a launch-on-warning posture, this would be a significant change in its nuclear posture. In the past, China’s nuclear units reportedly trained to launch a retaliatory strike days or perhaps even weeks after China suffered a nuclear attack. However, there are reports that China’s nuclear forces have reduced their response time to the order of hours or minutes. If a future Chinese launch-on-warning posture simply requires receiving warning of an incoming strike, this would represent a significant lowering of the threshold for nuclear use.

Perhaps, ironically, when Chinese strategists write about the risks of deliberate nuclear use in a crisis or conflict with the United States, they envision U.S. nuclear use. Shifts in the local conventional balance likely reduces the incentives for China to use nuclear weapons first. If the PLA can when the fight with only conventional weapons, it would presumably prefer to avoid the costs of nuclear use. However, those very shifts in the local conventional balance—along with the deteriorating political relationship—may increase the likelihood of China initiating large-scale military action in the first place, which introduces the conditions for nuclear use.

 

At the same time, China simply has more tools at its disposal with which to engage in deliberate nuclear use and the increased survivability of its strategic nuclear forces could reduced the perceived costs of nuclear escalation. For the moment, it appears that the no-first-use policy is still in place and official Chinese sources continue to reiterate it. However, while this kind of peacetime signaling can be an important reflection of intentions, plans can change in a crisis or conflict.

​

 

In The Logic of American Nuclear Strategy, Matthew Kroenig seems to argue that the size and composition of a nuclear arsenal provides overall negotiating advantage. Is China applying Kroenig’s point of view?

 

Unlikely. Recent work drawing on Chinese sources concludes that “there is little evidence that China views the nuclear balance as delicate or is adopting a competitive nuclear strategy. A more reactive approach to nuclear force building is not necessarily a more competitive approach.”

 

There is significant uncertainty about China’s nuclear future and many competing explanations for its nuclear behavior. The available evidence does not suggest that China agrees with the Kroenig thesis. I should also add that the empirical support for Kroenig’s argument has been significantly eroded by my own work and other research by scholars such as Todd Sechser and Matt Fuhrmann, Lauren Sukin and Abby Fanlo, and Kyungwon Suh.

 

There are some recent pieces by on the one hand, Alex Debs and, on the other, Benjamin Bahney and Braden Soper, relying on formal modeling to assess the relationship between the nuclear balance and success in nuclear crises which offer competing conclusions. However, the empirical record suggests that the balance does not matter much—rather, what matters may be how those forces are postured. Further, it can be somewhat difficult to draw generalizable conclusions about the impact of nuclear weapons.

 

As I frequently tell my students, deterrence occurs in the mind of the target. It may not matter what we think—it only matters what “they” (in this case, Chinese decision-makers) think.